Investigation of gene expression in patients with metastatic urothelial carcinoma

Aim

“Our aim is to investigate genes being significant up- or down regulated in patients with urothelial carcinoma depending on whether or not they have metastasis. Moreover we want to explore the importance of smoking in regards to the gene expression.”

Introduction: Bladder cancer

Approx. 500,000 cases worldwide in 2020.
Smoking is considered to account for half of cases.
Only 5% diagnosed with metastasis in distant body parts.

American Cancer Society. 2023. “What Is Bladder Cancer?” https://www.cancer.org/cancer/types/bladder-cancer/about/what-is-bladder-cancer.html.

Cancer.Net. 2023. “Bladder Cancer: Statistics” https://www.cancer.net/cancer-types/bladder-cancer/statistics

Data description

93 patients with urothelial carcinoma undergoing radical cystectomy
Microarray data from patients collected from 1993 to 2004
Only metadata we use is the smoking status and metastasis
Total number of genes is 54,675

	Overall (N=93)
Smoking
Current	19 (20.4%)
Former	56 (60.2%)
Never	18 (19.4%)
Metastasis
0	57 (61.3%)
1	36 (38.7%)

Riester M, Taylor JM, Feifer A, Koppie T et al. Combination of a novel gene expression signature with a clinical nomogram improves the prediction of survival in high-risk bladder cancer. Clin Cancer Res 2012 Mar 1;18(5):1323-33. PMID: 22228636

Methods

Data cleaning

Selecting smoking, metastasis, and gene expressions as variables
From wide to long format
Filtering genes with standard deviation greater than 0:
- From 54,675 to 48,932 genes

Methods

Augmentation

Analyze gene association between gene expression and metastasis.

Fold change
- Measure the magnitude of gene expression change.
Data augmentation: log2(fold change)
- Scale for better interpretability due to symmetry.
  
  \[ \text{Avg. log2 FC}_a = \log\left(\frac{\mu(\text{gene}_{\text{meta}})}{\mu(\text{gene}_{\text{non-meta}})}\right) \]
  
  Average Gene Expression (one value for each gene)
  
  \[ \text{Ind. log2 FC}_{\text{ind}} = \log\left(\frac{\text{gene}_{\text{meta}}}{\mu(\text{gene}_{\text{non-meta}})}\right) \]
  
  Individual Gene Expression (yes metastasis list for each gene)
T-TEST
- Compare the mean of two groups.
P-Value (cut)
- Set at 0.01.

Results

Aim of Analysis 1: Identification of Genes Exhibiting Significantly Different Expression between Metastatic and Non-Metastatic Cancer Patients.

Visualize significant genes and their regulation in a volcano plot
Using the average Log2 Fold Change for each gene
286 genes out of 48,932 genes are significantly different expressed in patients with metastasis on a significance level of 0.01

Riester M, Taylor JM, Feifer A, Koppie T et al. Combination of a novel gene expression signature with a clinical nomogram improves the prediction of survival in high-risk bladder cancer. Clin Cancer Res 2012 Mar 1;18(5):1323-33. PMID: 22228636

Results

Aim of Analysis 2A: Investigation of Altered Gene Expression in the most significant genes related to metastasis in bladder cancer.

Top 13 most significant genes were selected for further analysis
The figure shows the average Log2 Fold Change value, together with lower and upper 95% confidence intervals
5 genes are up-regulated, 8 genes are down-regulated

Riester M, Taylor JM, Feifer A, Koppie T et al. Combination of a novel gene expression signature with a clinical nomogram improves the prediction of survival in high-risk bladder cancer. Clin Cancer Res 2012 Mar 1;18(5):1323-33. PMID: 22228636

Results

Aim of analysis 2B: Investigation of the effect of smoking in relation to gene expression

Individual log2fold change for patients with metastasis
Count of Current Smokers, Former Smokers, and Non-smokers for each gene
13 most significant genes (p < 0.01) grouped based on up or down regulation being risk factor
Current smokers for all gene regulation of risk
More metastatic patients with genes expression being risk

Riester M, Taylor JM, Feifer A, Koppie T et al. Combination of a novel gene expression signature with a clinical nomogram improves the prediction of survival in high-risk bladder cancer. Clin Cancer Res 2012 Mar 1;18(5):1323-33. PMID: 22228636

Discussion

Out of 48,932 genes, 286 genes are significantly up or downregulated in patients with metastasis
Current smoking seem to have a relation to gene regulation inducing metastasis in patients with urothelial carcinoma
- Statistical test on relation to smoking

Small sample size (N = 93). More data will increase statistical power
Include more metadata to the analysis to expand the aim